Of the two million Americans with seizures disorders, approximately 300,000 suffer from intractable epilepsy and, of these, 50,000 are potential candidates for surgical treatment. The actual number of surgeries performed per year, however, is only 300-400. A major limitation is the difficulty of localizing the epileptogenic focus, a prerequisite for respective surgery. Currently, invasive electroencephalography (EEG) is often used to evaluate patients with seizure disorders. However, the placement of sub-dural electrode grids or depth electrodes is an expensive procedure which entails significant potential risks to the patient. While non-invasive techniques for mapping seizure foci have progressed rapidly over the last decade, particularly magnetic resonance imaging (MRI) and positron emission tomography (PET), the need for improved sensitivity and specificity remains. Proton magnetic resonance spectroscopic imaging (1HNMR) is a relatively new imaging modality which is just approaching technical maturity. Metabolic images can be produced with spatial resolution of the order of 1 cm3, using widely available high field MR scanners. Preliminary data from our group and others indicate that seizure foci are associated with metabolic changes which can be detected by 1H NMR, even in cases where conventional MRI scans appear normal. Our overall goal is to establish the role of combined magnetic resonance spectroscopy and MRI in the presurgical evaluation of patients with intractable with intractable epilepsy. 1H NMR and MRI may reduce the need for invasive monitoring procedures, assist in surgical planning, and provide new information regarding metabolic changes in human epilepsy. We will develop an optimum 1 HNMR protocol for the examination of epilepsy patients, in which our goal will be to implement quantitative chemical shift imaging techniques, maximize sensitivity, and improve image reconstruction algorithms. Combined 1H NMR and MRI will be performed in epilepsy patients who are candidates for respective surgery, and in an equal number of age- and sex-matched volunteers. Spectroscopic and imaging data will be compared to EEG findings, post-surgical outcome and the results of histological analysis.